Scalable Run Time Reconfigurable Architecture
Abdellah Touhafi, Wouter Brissinck and Erik Dirkx
Erasmus Hogeschool Brussel nijverheidskaai 1701070 Brussel Belgium
Vrije Universiteit Brussel Pleinlaan 2 1050 Brussel Belgium
Abstract: Currently multi-FPGA reeonfigurable eomputing systems are still eommonly
used for accelerating algorithms. This teehnology where acceleration is
aehieved by spatial implementation of an algorithm in reeonfigurable hardware
has proven to be feasible. However, the best suiting algorithms are those who
are very struetured, ean benefit from deep pipelining and need only loeal
eommunieation resourees. Many algorithms ean not fulfil the third requirement
onee the problem size grows and multi-FPGA systems beeome neeessary. In
this paper we address the emulation of a run time reeonfigurable proeessor
arehiteeture, whieh seales better for this kind of eomputing problems.
1. INTRODUCTION
Currently multi-Field Programmable Gate Array (FPGA) reconfigurable
computing systems are still commonly used for accelerating algorithms.
This technology where acceleration is achieved by spatial implementation of
an algorithm in reconfigurable hardware has proven to be feasible. However,
research pointed out that the application must fulfil some requirements in
order to achieve a high performance. The best suiting algorithms are those
who are very structured, can benefit from deep pipelining and need only
local communication resources. Many algorithms can not fulfil the third
requirement once the problem size grows and multi-FPGA systems become
L. M. Silveira et al. (eds.), VLSI: Systems on a Chip
© Springer Science+Business Media New York 2000